Language comprehension is not a passive act of receiving words. As the brain hears or reads a sentence, it continuously evaluates what has already arrived while estimating what is likely to come next. A new review in Nature Reviews Psychology argues that this predictive ability is far more flexible than a single, universal language mechanism. Across childhood, adulthood and older age, people appear to generate, select and use predictions in different ways, depending on their cognitive resources, linguistic experience and the immediate value of anticipating upcoming information.
In everyday communication, language is filled with uncertainty. Speakers may pause, change direction, use unfamiliar expressions or produce sounds in noisy environments. To keep up, the brain can construct a range of possible continuations. If someone says, “The child opened the…,” listeners may rapidly activate words such as “door,” “box” or “book,” even before the final word is spoken. These activated possibilities can make comprehension faster when the prediction is correct, but they can also create difficulty when the actual input violates expectations.
The review by Kara D. Federmeier, Hsin-Shan Chiu, Caitlin L. Fisher and colleagues emphasizes that prediction should not be treated as a single process that is either present or absent. It can involve generating detailed expectations about a specific word, selecting among several possible interpretations or preparing more general linguistic categories. The brain may anticipate a particular sound, grammatical structure or meaning, or it may simply become ready for information that is likely to be relevant. These forms of prediction can operate together, but they do not necessarily develop or decline in parallel.
This distinction is especially important when comparing people at different ages. Children are still building the vocabulary, grammatical knowledge and world knowledge needed to make precise predictions. Their expectations may therefore be less specific, not because their brains are incapable of prediction, but because they have fewer reliable linguistic patterns from which to construct it. At the same time, children may use the knowledge they do possess efficiently, gradually refining their predictions as they encounter more language.
Young adults often have substantial cognitive resources available for predictive processing. Working memory, attention and processing speed can help them maintain sentence context, compare competing interpretations and rapidly update expectations as new words arrive. Yet even in this period, prediction is not necessarily constant. People may engage it more strongly when the context is highly informative and when anticipating the next word offers a clear advantage. When the input is unpredictable or the cost of a mistaken expectation is high, the brain may rely more heavily on incoming sensory information.
Older adults present an equally complex picture. Age-related changes in processing speed, working memory or attentional control may make some forms of highly specific prediction more demanding. However, older adults often possess extensive vocabulary and accumulated language experience, which can support powerful expectations based on meaning, context and familiar patterns. This combination means that aging cannot be reduced to a simple loss of predictive ability. Instead, older adults may shift how prediction is deployed, relying on well-established knowledge while adjusting the level of specificity to match available resources.
The review frames these differences through the interaction of cognitive resources, linguistic experience and momentary utility. Cognitive resources determine how much information a person can maintain and manipulate while a sentence unfolds. Linguistic experience determines which patterns are familiar enough to support reliable expectations. Utility determines whether prediction is worth the effort in a particular situation. A listener in a quiet conversation may benefit from anticipating a speaker’s next phrase, while someone dealing with severe background noise may avoid overly specific guesses and wait for clearer evidence.
This flexible strategy also helps explain why prediction can sometimes improve comprehension and sometimes interfere with it. When an expectation matches the incoming word, processing may be more efficient because the brain has already prepared for that information. When the word is unexpected, the system must revise its interpretation, producing what researchers often describe as a prediction error. Such mismatches are not merely failures; they provide information that allows the brain to update its internal model of the sentence and the speaker’s intended meaning.
The authors’ developmental perspective challenges the idea that prediction is an invariant hallmark of language comprehension. Instead, prediction appears to be a dynamically configured system, with different components selectively engaged under different conditions. The same person may use detailed lexical predictions in one context, broad semantic expectations in another and minimal prediction when the language signal is unreliable. Across the lifespan, changes in prediction may therefore reflect adaptation as much as limitation.
Understanding this adaptability could influence research on language development, aging and communication disorders. It may also help explain why people who appear to have similar comprehension abilities can use different strategies to reach the same interpretation. Rather than asking whether the brain predicts, scientists may increasingly ask what it predicts, how specifically it predicts, when it chooses to predict and how it responds when expectations are wrong. The emerging picture is of language comprehension as a constantly recalibrated partnership between anticipation and evidence—one that changes with experience, capacity and the demands of the moment.
Subject of Research: Predictive mechanisms in language comprehension across the lifespan.
Article Title: Prediction in language comprehension across the lifespan.
Article References: Federmeier, K.D., Chiu, HS., Fisher, C.L. et al. Prediction in language comprehension across the lifespan. Nature Reviews Psychology (2026). https://doi.org/10.1038/s44159-026-00599-2
Image Credits: AI Generated
DOI: 10.1038/s44159-026-00599-2
Keywords: Language comprehension, prediction, predictive processing, lifespan development, cognitive resources, linguistic experience, children, young adults, older adults, prediction error.
